The Neural and Vascular Pathophysiology Presented in the Chronic Phase of Non-freezing Cold Injury.

  • Jennifer Anne Wright

Student thesis: Doctoral Thesis


Non-freezing cold injury (NFCI) may occur at the extremities when tissue cools to within the temperature range of 0 ºC to 15 ºC, but remain above freezing. This is following continuous or repeated exposure to cold and often wet conditions. NFCI has been noted throughout history and is often thought of in its most extreme form in which gangrene, muscle wastage, prolonged pain, and permanent disability can occur. Whilst this was prevalent as recently as the First and Second World Wars, the presentation of NFCI is now generally more subtle. Despite this, the injury can still be debilitating for patients as sensory changes, neuropathic pain, and cold sensitivity can persist for years, and potentially a lifetime. Although the shift in the phenotype of the injury has been documented in literature, understanding of the pathophysiology has remained limited. The research presented often uses differing methodologies in NFCI patient groups, and measures are not compared to rigorously standardised control data. Consequently, this has led to ambiguity in the diagnosis and effective treatment of NFCI.
To address this need, Study 1 investigated NFCI in the chronic phase in patients (NFCI) who were all serving members of the British Army. Data collection included vascular, plasma biomarker, neural and questionnaire measures. The findings of these measures were compared with two control groups. The first being a cold exposed control group (COLD-CON), which consisted of British Army personnel who had undertaken similar training and cold exposure to the injured cohort, but did not have NFCI. The second control group (CON) was a non-cold-exposed group, who were not members of the British armed forces, and did not report experiencing great levels of cold exposure in the two years prior to participation in the study. The three groups were matched for age, sex, height, mass, ethnicity, estimated aerobic fitness, and foot volume, as all of these factors may contribute towards individual variation in neural and vascular function. Primary findings indicated that perceptual measures differentiated the NFCI from the control groups, with local thermal sensation being colder, thermal comfort being more uncomfortable, and greater pain felt in NFCI patients in the cold sensitivity test and foot cooling protocols. The range of questionnaires implemented indicated that the NFCI had a greater inability to cope in cold environments when compared to the control groups, and this was accompanied by pain which affected the patients’ ability to carry out day to day tasks.
Few physiological measures differentiated between the patient and control groups, and where differences were present, they encompassed a range of vascular, plasma biomarker, and neural mechanisms. These included a reduced inspiratory gasp vascular response in 30 ºC in the NFCI patients, and a lower mean toe temperature from the cold sensitivity test in NFCI patients. 4-Hydroxynonenal following both hot and cold thermal challenges was lower in NFCI patients, whereas interleukin-10 (IL-10) measured at baseline was higher in NFCI. The mechanical and warm detection thresholds from quantitative sensory testing were higher in NFCI. Intraepidermal nerve fibre density in unmatched groups was lower in NFCI patients. Although these differences in perceptual and physiological measures were highlighted from Study 1, a high degree of variability in the data presented between the three groups was evident. Consequently, this limits the clinical utility of each measure as a diagnostic indicator for NFCI, owing to the likelihood of Type 1 or 2 errors.
In an attempt to confirm the pathophysiological mechanisms involved in NFCI, and assess the clinical appropriateness of key measures, 16 assessments were carried forward from Study 1 to Study 2. In this, a more homogenous NFCI participant group was recruited from the British Army, where all patients had been diagnosed with NFCI within a lower mean timeframe from the onset of injury to participation in the study. This was implemented in an attempt to minimise variation in the results, which may have been caused by differences in the duration of the sustained injury. To assess this, the NFCI patients in Study 2 were compared with the NFCI patients in Study 1. Following this, the data from the NFCI patients in Study 2 were added to the data from the NFCI patients in Study 1, and compared to all data available from the COLD-CON and CON control groups. From the additional data sets, it was hoped this would reduce the variability in the data presented, and increase the statistical power of the comparisons made between groups. As such, it was hypothesised that when the data from the NFCI group in Study 2 is combined with the data from the NFCI group from Study 1, the difference between the NFCI patients and control groups that were present in Study 1 (COLD and CON) will be confirmed by the additional data. This would then determine which tests are clinically robust enough to determine key features in the pathophysiology of NFCI, and which tests may aid in forming a clinical diagnosis. However, this analysis revealed only two
physiological tests which remained different between NFCI and control participants (the mechanical and warm detection thresholds remained higher in NFCI patients). This led to the conclusion that different forms of analysis needed to be explored to better interpret the variable data, as it may be that the pathophysiology of NFCI in the patient cohort was masked in the between group analysis by extraneous variables.
Following on from Study 2, the variability in the data continued to prevent a comprehensive understanding of the pathophysiology of NFCI from being obtained. This impacted upon our understanding of how NFCI might be diagnosed. In an attempt to address this, Study 3 utilised the data collected from Study 1 and 2 to conduct four different forms of data analysis (the dominant rough set approach, cluster analysis, ordinal logistic regression analysis, and Z-score analysis) with the aim to address: (i) if common themes in the pathophysiology of NFCI are present, and (ii) how they may be used to form a diagnosis. It was concluded, that all four analyses have the potential to aid in our understanding of the pathophysiology of NFCI and the formation of a diagnosis. The clustering analysis and ordinal logic regression analysis confirmed that altered measures of mechanical detection and warm detection (which may indicate impaired neural function in unmyelinated C and A-delta, and A-Beta nerve fibres, as highlighted in Study 2), are likely common mechanisms that are impaired in NFCI patients. Whereas the Z-score analysis indicated that although these may be commonly implicated mechanisms, it is likely that a range of mechanisms may be impaired in different patients resulting from situational differences at the time of injury. Finally, the dominant rough set analysis suggested that the perceptual measures (questionnaires, thermal sensation, thermal comfort, and pain) are the strongest to differentiate between NFCI patients and control groups. However, using these measures within a diagnostic context can be problematic with objective measures preferred where employment and compensation are involved.
Owing to these conflicting outcomes, it is suggested that each form of analysis requires additional investigation to improve their effectiveness, before they can be successfully implemented within a clinical context. This will thereby reduce the risk of a Type 2 error when diagnosing future patients with NFCI.
Date of Award14 Sept 2023
Original languageEnglish
Awarding Institution
  • University of Portsmouth
SupervisorClare Eglin (Supervisor), Heather Massey (Supervisor) & Mike Tipton (Supervisor)

Cite this